An unmanned aerial vehicle cluster cooperative navigation positioning method, system, device and medium

Abstract

The application relates to the technical field of navigation, and discloses a method, system, device and medium for cooperative navigation and positioning of a UAV cluster. The method comprises the following steps: acquiring theoretical covariance and actual covariance when UKF is used for cooperative navigation and positioning, and adjusting the theoretical noise variance in the theoretical covariance, so that the theoretical covariance after the adjustment of the theoretical noise variance is equal to the actual covariance; extracting relative measurement data of a relative navigation sensor in a cooperative navigation system of the UAV cluster; capturing a dependency relationship between a relative position calculation value and a relative position measurement value in the relative measurement data, predicting the relative position measurement value through the dependency relationship and the relative position calculation value, and reconstructing the relative measurement data by combining the predicted relative position measurement value and the relative position calculation value; and performing cooperative navigation and positioning of the UAV cluster according to the UKF after the adjustment of the theoretical noise variance or making the UKF perform cooperative navigation and positioning of the UAV cluster according to the reconstructed measurement data, so that the accuracy is relatively high.

Description

Technical Field

[0001] This invention relates to the field of navigation technology, and in particular to a method, system, device and medium for collaborative navigation and positioning of unmanned aerial vehicle (UAV) swarms. Background Technology

[0002] With the development of artificial intelligence and swarm technology, drone swarms have been widely used in both civilian and military fields. Swarm formations expand the execution capabilities and mission range of individual drones through mutual cooperation, and drone swarm cooperative navigation technology is the prerequisite and core for realizing formation control and intelligent operation, thus becoming one of the research hotspots.

[0003] Cooperative navigation in drone swarms primarily employs two modes: parallel and master-slave. The former requires all drones to possess high-precision navigation capabilities, thus limiting its applicability to small-scale swarms. In contrast, the master-slave mode is relatively simpler, requiring only a small number of drones with high-precision navigation capabilities as the "leader," while the "wingmen" acquire navigation information relative to the "leader" through low-cost sensors to achieve self-localization. Due to its simplicity, flexibility, and low cost, the master-slave mode is more suitable for practical applications.

[0004] However, low-cost sensors used in cooperative navigation, such as Miniature Inertial Measurement Units (MIMUs) and other relative navigation sensors (such as laser angle and range sensors and Doppler velocity sensors), all have their limitations. MIMUs, due to significant inertial sensor drift, experience rapid accumulation of navigation errors over time, while relative navigation sensors are susceptible to interference from complex environments and maneuvering conditions, thus affecting the cooperative navigation performance of UAVs. To address these limitations, integrated navigation provides an effective means to achieve complementary advantages among these sensors.

[0005] Data fusion is a key technology for realizing integrated navigation, processing measurement data from MIMUs and relative navigation sensors to provide accurate cooperative navigation information. Currently, Kalman filtering (KF) is the core method for achieving multi-sensor cooperative navigation. However, due to the error characteristics of MIMUs and the high maneuverability of UAVs, cooperative navigation models often exhibit nonlinear characteristics. Therefore, Kalman filtering, which is only applicable to linear systems, cannot be directly applied to nonlinear cooperative navigation. To address this, researchers have focused on studying various nonlinear filtering methods, such as extended Kalman filtering (EKF) and unscented Kalman filtering (UKF).

[0006] The EKF approximates the linearization of a nonlinear system through Taylor series expansion and ignores higher-order terms. However, for strongly nonlinear systems, the EKF linearization error is relatively large, and the calculation of the Jacobian matrix is ​​quite complex. The UKF generates a set of Sigma points through an unscented transform (UT) to approximate the mean and covariance of the system state, avoiding the calculation of the Jacobian matrix, and achieving higher state estimation accuracy than the EKF. Therefore, the UKF becomes a superior method for solving the state estimation problem of nonlinear systems.

[0007] Due to the inherent characteristics of Kalman filtering, UKF-based data fusion requires accurate measurement models and their noise statistics. If the measurement model and its noise statistics are flawed, the state estimation results will be affected, potentially leading to filter divergence. However, as mentioned earlier, relative navigation sensors are susceptible to interference from complex environments and maneuvering conditions, resulting in abnormal measurements (e.g., measurement loss due to sensor malfunction or obstruction, measurement anomalies / drifts caused by sensor malfunction or maneuvering, and changes / deviations in measurement noise statistics due to environmental interference), making accurate cooperative navigation positioning impossible. Summary of the Invention

[0008] The purpose of this invention is to provide a method, system, device and medium for collaborative navigation and positioning of unmanned aerial vehicle (UAV) swarms, which can solve the technical problem that relative navigation sensors are easily affected by complex environments and maneuvering states, resulting in abnormal measurements and the inability to achieve accurate collaborative navigation and positioning.

[0009] To address the aforementioned technical problems, embodiments of the present invention provide a method for cooperative navigation and positioning of unmanned aerial vehicle (UAV) swarms, comprising the following steps:

[0010] Establish a measurement model for a UAV swarm collaborative navigation system to obtain collaborative measurement data of the UAV swarm obtained through all navigation sensors in the UAV swarm collaborative navigation system;

[0011] The theoretical and actual innovation covariance of the unscented Kalman filter (UKF) during cooperative navigation and positioning based on cooperative measurement data are obtained. The theoretical measurement noise variance in the theoretical innovation covariance is adjusted so that the theoretical innovation covariance after adjusting the theoretical measurement noise variance is equal to the actual innovation covariance.

[0012] The relative measurement data is extracted from the cooperative measurement data of the UAV swarm cooperative navigation system relative to the navigation sensor; wherein, the relative measurement data consists of the calculated relative position / velocity between the leader and wingmen in the UAV swarm and the relative position / velocity measurement values ​​between the leader and wingmen measured by the relative navigation sensor;

[0013] Capture the dependency between the calculated and measured relative position / velocity values ​​in the relative measurement data, and predict the measured relative position / velocity values ​​using the dependency and the calculated values, so as to reconstruct the relative measurement data by combining the predicted and calculated values.

[0014] The UKF is used to perform collaborative navigation and positioning of UAV swarms based on the adjusted theoretical measurement noise variance, or the UKF is used to perform collaborative navigation and positioning of UAV swarms based on the reconstructed measurement data.

[0015] Optionally, adjusting the theoretical measurement noise variance in the theoretical innovation covariance to make the adjusted theoretical innovation covariance equal to the actual innovation covariance includes:

[0016] The theoretical information covariance and the actual information covariance are input into the first adaptive neurofuzzy inference network ANFIS;

[0017] The matching degree between the theoretical innovation covariance and the actual innovation covariance is obtained through the ANFIS network, and an adjustment coefficient is output based on the matching degree to adjust the theoretical measurement noise variance so that the theoretical innovation covariance after adjusting the theoretical measurement noise variance is equal to the actual innovation covariance.

[0018] Optionally, the step of capturing the dependency between the calculated relative position / velocity value and the measured relative position / velocity value in the relative measurement data, and predicting the measured relative position / velocity value through the dependency and the calculated relative position / velocity value, includes:

[0019] Using the calculated and measured relative positions / velocities of the lead and wingmen of several drone swarms as sample data, a second ANFIS network is trained to obtain an ANFIS model, which captures the dependency between the calculated and measured relative positions / velocities.

[0020] The calculated relative position / velocity values ​​are input into the ANFIS model to obtain the predicted relative position / velocity measurements.

[0021] Optionally, before obtaining the theoretical and actual innovation covariance of the unscented Kalman filter (UKF) during cooperative navigation and positioning based on cooperative measurement data, the method further includes:

[0022] Construct an anomaly detection function for collaborative measurement data;

[0023] The anomaly detection function is:

[0024] ;

[0025] In the formula, It is a unit vector, and the first... Each element is 1. It is the dimension of the measurement vector. For the innovation vector, This is the corresponding covariance matrix;

[0026] Based on the anomaly detection function, set the first event trigger condition, and determine whether the collaborative measurement data is abnormal based on the first event trigger condition;

[0027] The trigger condition for the first event is:

[0028] ;

[0029] In the formula, It is the false alarm rate. It is the detection threshold.

[0030] Optionally, before obtaining the theoretical and actual innovation covariance of the unscented Kalman filter (UKF) during cooperative navigation and positioning based on cooperative measurement data, the method further includes:

[0031] If the collaborative measurement data is determined to be abnormal, then a relative position change index of the collaborative measurement data is constructed.

[0032] The relative position change index is:

[0033] ;

[0034] In the formula, This indicates the measurement dimension involved in the anomaly. The total dimension of the measurement. It is during the quarantine period After the dimensional measurement, the 3rd dimension of the state estimation error covariance matrix diagonal elements, During UKF filtering, the state estimation error covariance matrix is ​​the first... One diagonal element;

[0035] Based on the relative position change index, set the second event trigger condition, and based on the second event trigger condition, determine whether to adjust the theoretical measurement noise variance or whether to predict the relative position / velocity measurement value.

[0036] The trigger condition for the second event is:

[0037] ;

[0038] In the formula, It is the threshold for triggering the second event.

[0039] Embodiments of the present invention also provide a collaborative navigation and positioning system for unmanned aerial vehicle (UAV) swarms, comprising:

[0040] The model building module is used to build a measurement model of the UAV swarm collaborative navigation system in order to obtain the collaborative measurement data of the UAV swarm obtained by all navigation sensors in the UAV swarm collaborative navigation system.

[0041] The noise adjustment module is used to obtain the theoretical and actual innovation covariance of the unscented Kalman filter UKF when performing cooperative navigation and positioning based on cooperative measurement data, and to adjust the theoretical measurement noise variance in the theoretical innovation covariance so that the theoretical innovation covariance after adjusting the theoretical measurement noise variance is equal to the actual innovation covariance.

[0042] The data acquisition module is used to extract relative measurement data from the collaborative measurement data of the UAV swarm collaborative navigation system relative to the navigation sensor; wherein, the relative measurement data consists of the calculated relative position / velocity between the leader and wingmen in the UAV swarm and the relative position / velocity measurement values ​​between the leader and wingmen measured by the relative navigation sensor;

[0043] The position / velocity prediction module is used to capture the dependency between the calculated and measured values ​​of relative position / velocity in the relative measurement data, and predict the measured values ​​of relative position / velocity through the dependency and the calculated values ​​of relative position / velocity, so as to reconstruct the relative measurement data by combining the predicted measured values ​​of relative position / velocity and the calculated values ​​of relative position / velocity.

[0044] The navigation and positioning module is used to perform collaborative navigation and positioning of UAV swarms based on the UKF after adjusting the theoretical measurement noise variance, or to enable the UKF to perform collaborative navigation and positioning of UAV swarms based on the reconstructed measurement data.

[0045] Embodiments of the present invention also provide a computer device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the above-described UAV swarm cooperative navigation and positioning method.

[0046] Embodiments of the present invention also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described UAV swarm cooperative navigation and positioning method.

[0047] The UAV swarm cooperative navigation and positioning method provided by this invention has at least the following beneficial effects:

[0048] This invention reduces the negative impact of abnormal measurement values ​​in UAV swarm cooperative navigation systems through the following two methods, thereby improving the accuracy of subsequent UAV swarm cooperative navigation and positioning by UKF:

[0049] One approach involves acquiring the theoretical and actual innovation covariance of the Unscented Kalman Filter (UKF) during cooperative navigation and positioning based on cooperative measurement data. The theoretical measurement noise variance within the theoretical innovation covariance is then adjusted to ensure that the adjusted theoretical innovation covariance is equal to the actual innovation covariance. Since the UKF theoretical innovation covariance includes the theoretical measurement noise variance from the cooperative navigation system, the closer the theoretical measurement noise variance is to the actual measurement noise variance, the closer the theoretical innovation covariance is to the actual innovation covariance. This method can reduce measurement noise bias and mitigate the impact of abnormal measurements.

[0050] Another approach involves extracting relative measurement data from the drone swarm's cooperative navigation system relative to the navigation sensors. This captures the dependency between the calculated and measured relative position / velocity values ​​within the relative measurement data. The relative position / velocity measurements are then predicted based on this dependency and the calculated values. Finally, the relative measurement data is reconstructed by combining the predicted and calculated values. Since the relative measurement data consists of the calculated relative position / velocity between the lead and wingmen in the drone swarm, as well as the measured values ​​from the navigation sensors, and the calculated values ​​are relatively fixed, it can be assumed that most measurement anomalies are due to the relative position / velocity measurements. By introducing measurement prediction to reconstruct the measurement data, the impact of abnormal measurements can be reduced. Attached Figure Description

[0051] One or more embodiments are illustrated by way of example with reference to the accompanying drawings, and these illustrative descriptions do not constitute a limitation on the embodiments.

[0052] Figure 1 This is a flowchart of a collaborative navigation and positioning method for unmanned aerial vehicle (UAV) swarms according to an embodiment of the present invention;

[0053] Figure 2 This is a schematic diagram of relative navigation measurement for a wingman according to an embodiment of the present invention;

[0054] Figure 3 This is a schematic diagram of a dual-event-triggered ANFIS mode structure according to an embodiment of the present invention;

[0055] Figure 4This is a schematic diagram of an ANFIS structure and optimization process according to an embodiment of the present invention;

[0056] Figure 5 This is a schematic diagram of the flight trajectory of a drone swarm provided according to an embodiment of the present invention;

[0057] Figure 6 This is a schematic diagram of the root mean square error curves of three algorithms under the condition of abrupt measurement changes, according to an embodiment of the present invention.

[0058] Figure 7 This is a schematic diagram of the cumulative root mean square error of position under the condition of a sudden change in measurement, provided by one embodiment of the present invention;

[0059] Figure 8 This is a schematic diagram of the root mean square error curves of three algorithms under non-Gaussian measurement noise conditions according to an embodiment of the present invention;

[0060] Figure 9 This is a schematic diagram illustrating the cumulative root mean square error of position under non-Gaussian measurement noise using three algorithms according to an embodiment of the present invention;

[0061] Figure 10 This is a schematic diagram of the root mean square error curves of three algorithms under the condition of measurement loss, according to an embodiment of the present invention;

[0062] Figure 11 This is a schematic diagram illustrating the cumulative root mean square error of location using three algorithms in the case of measurement loss, according to an embodiment of the present invention.

[0063] Figure 12 This is a schematic diagram comparing the computational efficiency of UKF, UKF-ANFIS, and UKF-DEANFIS according to an embodiment of the present invention. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details are presented in the embodiments of the present invention to facilitate a better understanding of the invention. However, the technical solutions claimed in the present invention can be implemented even without these technical details and various variations and modifications based on the following embodiments. The division of the following embodiments is for ease of description and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined with and referenced by each other without contradiction.

[0065] One embodiment of the present invention relates to a method for cooperative navigation and positioning of unmanned aerial vehicle (UAV) swarms. The specific process of the UAV swarm cooperative navigation and positioning method in this embodiment can be as follows: Figure 1 As shown, it includes:

[0066] Step 101: Establish a measurement model for the UAV swarm collaborative navigation system to obtain collaborative measurement data of the UAV swarm obtained through all navigation sensors in the UAV swarm collaborative navigation system.

[0067] Step 102: Obtain the theoretical and actual innovation covariance of the unscented Kalman filter (UKF) when performing cooperative navigation and positioning based on cooperative measurement data, and adjust the theoretical measurement noise variance in the theoretical innovation covariance so that the theoretical innovation covariance after adjusting the theoretical measurement noise variance is equal to the actual innovation covariance.

[0068] Step 103: Extract relative measurement data from the relative navigation sensor in the UAV swarm cooperative navigation system from the cooperative measurement data; wherein, the relative measurement data consists of the calculated relative position / velocity between the leader and wingmen in the UAV swarm and the relative position / velocity measurement values ​​between the leader and wingmen measured by the relative navigation sensor.

[0069] Step 104: Capture the dependency between the calculated relative position / velocity and the measured relative position / velocity in the relative measurement data, and predict the measured relative position / velocity using the dependency and the calculated relative position / velocity, so as to reconstruct the relative measurement data by combining the predicted measured relative position / velocity and the calculated relative position / velocity.

[0070] The following is a detailed description of the implementation details of the UAV swarm cooperative navigation and positioning method in this embodiment. The following content is only for the convenience of understanding and is not necessary for implementing this solution.

[0071] Step 1: Establish a measurement model for the UAV swarm collaborative navigation system to obtain collaborative measurement data of the UAV swarm obtained through all navigation sensors in the UAV swarm collaborative navigation system.

[0072] In the specific implementation, a state model of the collaborative navigation system is established based on the error equation of the wingman MIMU in the UAV swarm collaborative navigation system; relative navigation vector measurement values ​​and relative navigation vector calculation values ​​are obtained based on the relative distance, speed and angle information obtained by the MIMU of the lead aircraft and the wingman as well as the relative navigation sensor, and a measurement model of the collaborative navigation system is established by subtraction.

[0073] (1) Based on the error equation of MIMU, establish the state model of the cooperative navigation system:

[0074] To establish a cluster-based collaborative navigation model, the following coordinate system is used: inertial coordinate system ( (system), carrier coordinate system ( Earth coordinate system (system) and navigation coordinate system ( Tie), The system is usually chosen as the East-North-Sky geographic coordinate system ( Furthermore, due to rotational errors, the calculated navigation coordinate system differs from the actual coordinate system. There are differences in the system, which are defined as follows: System. The wingman MIMU error equation is selected as the state model for cluster cooperative navigation, where the state vector is:

[0075] ;

[0076] In the formula, , and These are the attitude, velocity, and position errors of the MIMU; , These represent the constant drift of the gyroscope and accelerometer, respectively.

[0077] Therefore, the MIMU attitude error equation can be written as:

[0078] ;

[0079] In the formula, and From System and Tie The transformation matrix of the system; This represents the transformation matrix that represents the relationship between the angular velocity of the computational platform coordinate system and the error angle of the Euler platform. yes System relative to The angular velocity of the system, That is the corresponding error.

[0080] Furthermore, the MIMU velocity and position errors can be written as:

[0081] ;

[0082] ;

[0083] In the formula, It is the specific force measured by the accelerometer. It is the Earth's angular velocity. It is the angular velocity relative to the Earth. and yes and The error, A and B It is a coefficient matrix, specifically represented as:

[0084]

[0085] In the formula, and These are the wingman's speeds traveling east and north, respectively. and These are the principal radii of curvature along the Earth's meridian and circumference, respectively. and The latitude and altitude of the wingman.

[0086] MIMU error also includes the output errors of the gyroscope and accelerometer, which can be described as:

[0087] ;

[0088] In the formula, and It corresponds to white noise.

[0089] The state model of the wingman's cooperative navigation system can be obtained and further discretized as follows:

[0090] ;

[0091] In the formula, yes The state vector at any given time; It is a discrete nonlinear function describing the state model. It is zero-mean Gaussian white noise.

[0092] (2) Establish a measurement model for the cooperative navigation system:

[0093] according to Figure 2 The relative distance between the lead and wingman shown can be seen in the wingman's... The system is decomposed into:

[0094] ;

[0095] In the formula, and It is the relative distance between the lead aircraft and the wingman. exist Components in the system and These are the measured values ​​of relative azimuth and elevation angles.

[0096] Because the measured values ​​include errors, therefore... , and Considering , and If the value is smaller, the above expression can be further reformulated as:

[0097] ;

[0098] In the formula, values ​​with the symbol "~" represent measured values, while values ​​without "~" represent true values; This indicates measurement error.

[0099] Furthermore, The relative position vector measurement in the system can be expressed as:

[0100] ;

[0101] Furthermore, relative position vectors can be constructed based on the MIMU to calculate values. The lead aircraft and wingman... The position in the system can be In the system, it is represented as:

[0102] ;

[0103] In the formula, and These indicate the positions of the lead and wingman, respectively. It is the Earth's eccentricity.

[0104] The relative positional error between the lead aircraft and the wingman can be expressed as:

[0105] ;

[0106] The relative positions between the lead aircraft and the wingman can be calculated as follows:

[0107] ;

[0108] In the formula, From Tie The transformation matrix can be calculated using the position information output by the wingman's MIMU.

[0109] Finally, the relative position error vector can be obtained.

[0110] ;

[0111] Similar to the process described above, the relative velocity error vector can also be established as follows:

[0112] ;

[0113] The measurement model for wingman cooperative navigation can be obtained as follows:

[0114] ;

[0115] In the formula, It is the overall measurement noise. It is a measurement vector.

[0116] Step 2: Design a dual-event triggering mechanism to reduce unnecessary computation when dealing with normal measurements and improve the computational performance of ANFIS in cluster cooperative navigation. First, construct an anomaly detection function for cooperative measurement data. Based on the anomaly detection function, set the first event trigger condition and determine whether the cooperative measurement data is abnormal. Second, if the cooperative measurement data is determined to be abnormal, construct the relative position change index of the cooperative measurement data. Based on the relative position change index, set the second event trigger condition and determine whether to adjust the variance of theoretical measurement noise or whether to predict the relative position / velocity measurement value.

[0117] In its implementation, the dual-event triggering mechanism structure is as follows: Figure 3 As shown, there are two event triggering conditions:

[0118] Event Trigger Condition 1 (i.e., the first event trigger condition): This condition aims to detect measurement faults or anomalies to avoid unnecessary predictions and optimizations in the absence of faults or anomalies. This condition constructs a chi-square test based on the innovation vector of the filter, where the innovation vector is defined as:

[0119] ;

[0120] In the absence of measurement anomalies, the above formula should follow a zero-mean Gaussian distribution, and its error covariance matrix should be calculated as follows:

[0121] ;

[0122] When a measurement anomaly occurs, the mean of the innovation vector will no longer be zero. Therefore, an anomaly detection function can be constructed as follows:

[0123] ;

[0124] In the formula, It is a unit vector, and its first... Each element is 1. It is the dimension of the measurement vector. For the innovation vector, Let be the corresponding covariance matrix.

[0125] Therefore, the corresponding event triggering condition is defined as:

[0126] ;

[0127] In the formula, It is the false alarm rate. That is the corresponding detection threshold.

[0128] Event triggering condition 2 (i.e., the second event triggering condition): If the innovation vector does not satisfy triggering condition 1, i.e., an anomaly exists, then the decision to isolate the corresponding measurement is made based on the degree of impact of the measurement on the filtering accuracy. Therefore, in order to evaluate the impact of abnormal measurements on filtering accuracy, this invention constructs a relative position change index (RPCI) to determine whether ANFIS should be used for network prediction.

[0129] RPCI is defined as:

[0130] ;

[0131] In the formula, This indicates the measurement dimension involved in the anomaly. The total dimension of the measurement. It is during the quarantine period After the dimensional measurement, the 3rd dimension of the state estimation error covariance matrix diagonal elements, During UKF filtering, the state estimation error covariance matrix is ​​the first... One diagonal element.

[0132] The above formula reflects the degree of decrease in filtering accuracy after isolating a certain measurement. A larger RPCI value indicates a greater impact of that measurement on filtering accuracy. Therefore, the second event trigger condition can be constructed as follows:

[0133] ;

[0134] In the formula, It is the threshold for the second event to be triggered.

[0135] Therefore, the event triggering mechanism is established as follows:

[0136] ;

[0137] As can be seen from the above formula, if... This indicates that there are no abnormal measurements in the cooperative navigation, and a standard UKF can be used. If... This indicates that the sensor measurements have little impact on the filtering accuracy, and the sensor can be directly isolated. In this case, the UKF time update results are used as the final system state estimate. If If the value is 0, it indicates the presence of an abnormal measurement that is critical to the filtering result. Therefore, this measurement should be retained, and ANFIS should be used to further compensate for any errors caused by its presence.

[0138] Step 3: Design prediction and optimization adjustment strategies for an Adaptive Neuro-Fuzzy Inference System (ANFIS). The ANFIS structure and optimization process are as follows: Figure 4 As shown. Based on ANFIS, this invention addresses the negative impact of measurement anomalies in two ways: one is by adjusting the measurement noise covariance to reduce the weight of anomaly measurements, and the other is by introducing measurement prediction values.

[0139] There are two methods to use adaptive neurofuzzy inference systems to handle the negative impact of outlier measurements: one is to adjust the measurement noise variance matrix to reduce the weight of outlier measurements; the other is to introduce prediction to assist the filtering process.

[0140] (1) Measurement noise variance adaptive adjustment: Obtain the theoretical and actual information covariance of the unscented Kalman filter UKF when performing cooperative navigation positioning based on cooperative measurement data, and adjust the theoretical measurement noise variance in the theoretical information covariance so that the theoretical information covariance after adjusting the theoretical measurement noise variance is equal to the actual information covariance.

[0141] In practical implementation, the theoretical innovation covariance of the system is defined. variance of the actual innovation sequence for:

[0142] ;

[0143] ;

[0144] At each time step At this point, if the theoretically measured noise variance The closer the value is to the actual measurement noise variance, the higher the theoretical information covariance. The closer to the actual information covariance However, in practice, sensor malfunctions can cause... and The differences are due to adjustments needed to improve UKF performance. To reduce and The difference between them is called covariance matching.

[0145] Therefore, the simplest method is to make the two covariances equal, and then update... It can be calculated using the following formula:

[0146] ;

[0147] However, in the actual calculation process of data fusion algorithms, direct subtraction may result in a negative definite value, causing filtering to fail. To solve this problem, the theoretical and actual information covariances are input into the first adaptive neural fuzzy inference network (ANFIS). The ANFIS network obtains the matching degree between the theoretical and actual information covariances, and outputs an adjustment coefficient based on the matching degree to adjust the variance of the theoretical measurement noise, so that the theoretical information covariance after adjusting the variance of the theoretical measurement noise is equal to the actual information covariance.

[0148] In the specific implementation, an adaptive estimation algorithm based on fuzzy logic is designed to avoid potential systematic errors and enhance the robustness of the system. The designed fuzzy system is based on... and As input, it is represented as the degree of matching (DOM):

[0149] ;

[0150] Based on the DOM, an algorithm based on fuzzy logic outputs an adjustment coefficient. Adjust the measurement noise variance .Will and The matching degree between the two is used as the input to the fuzzy logic system. Then, the UKF measurement noise variance is adjusted. To make and Matching, that is:

[0151] ;

[0152] Under normal circumstances, There are three cases with the DOM:

[0153] a) If , Compare Large, then ,Increase To reduce the difference between the two covariances;

[0154] b) If , and If they are similar, then ,Keep constant;

[0155] c) If , Compare Small, then , reduce To reduce the difference between the two covariances.

[0156] Based on the above relationships, the following fuzzy rules can be defined:

[0157] Rule 1. if , then Increase;

[0158] Rule 2. if , then Remain unchanged;

[0159] Rule 3. if , then Decrease.

[0160] in, and These are two small thresholds used to distinguish different fuzzy rules.

[0161] Furthermore, the membership function of the input can be defined based on the scope of the DOM as follows:

[0162] Large:

[0163] ;

[0164] Equal:

[0165] ;

[0166] Small:

[0167] ;

[0168] In the formula, the parameter , and Used to determine the shape of the input membership function.

[0169] Based on the fuzzy rules established above, it is possible to... The range determines the output membership function as follows:

[0170] Large:

[0171] ;

[0172] Equal:

[0173] ;

[0174] Small:

[0175] ;

[0176] In the formula, the parameter , , and Used to determine the shape of the output membership function. It is worth noting the parameters of the input and output membership functions ( and It is obtained by training the ANFIS network.

[0177] Finally, the center-of-mass method was used to... Perform deblurring:

[0178] ;

[0179] (2) Measurement Prediction: Relative measurement data from the navigation sensors in the UAV swarm collaborative navigation system is extracted from the collaborative measurement data. This relative measurement data consists of calculated relative position / velocity values ​​between the lead and wingmen in the UAV swarm, and measured relative position / velocity values ​​between the lead and wingmen as measured by the navigation sensors. The dependency relationship between the calculated and measured relative position / velocity values ​​is captured, and the relative position / velocity values ​​are predicted using this dependency relationship and the calculated values. The predicted and calculated relative position / velocity values ​​are then combined to reconstruct the relative measurement data. The flight trajectories of the lead and wingmen in the UAV swarm are shown below. Figure 5 As shown.

[0180] In practical implementation, since the measurement values ​​in the measurement model are composed of... / and / Composition. Abnormal measurements indicate... / An anomaly has occurred. In this situation, an adaptive neurofuzzy inference system can be used to capture it. / and / ANFIS determines the dependencies between these dependencies to obtain predicted values. This process consists of two phases: a training phase and a prediction phase. During the training phase, ANFIS uses... / and / The network is trained using the original input data to construct the ANFIS model. Notably, the ANFIS model, along with its associated fuzzy membership functions and fuzzy rules, is built through learning from a large amount of data collected under normal measurement conditions. During the prediction phase, [the following will be used]. / The data is fed into the prediction model and the network model trained during the training phase is used to obtain... / The predicted value, i.e. / Based on the predicted values, the measured values ​​can be predicted and reconstructed to obtain... ,Right now:

[0181]

[0182] Simulation verification:

[0183] To better evaluate the robustness and effectiveness of the proposed UKF-DEANFIS in the presence of anomalous measurements, simulations were performed to evaluate it, and it was compared with the traditional UKF algorithm and the UKF algorithm based on adaptive neural network fuzzy inference (UKF-ANFIS). Three typical scenarios were considered in the simulations: i) anomalous measurement mutations; ii) non-Gaussian measurement noise; and iii) missing measurements. Furthermore, 20 Monte Carlo simulations were conducted, and the root mean square error (RMSE) was used to evaluate the performance of the three algorithms.

[0184] ;

[0185] In the formula, This represents the total number of Monte Carlo simulations. Indicates the first The navigation parameter error vector of the wingman in the sub-Monte Carlo simulation; express The Euclidean norm.

[0186] Furthermore, to further evaluate the performance of the three algorithms mentioned above, the cumulative root mean square error (ARMSE) of the wingman's position is calculated, as defined below:

[0187] ;

[0188] In the formula, Indicates the number of data points used for comparison.

[0189] Simulation parameter settings:

[0190] The sensor parameters required for wingman integrated navigation are shown in Table 1:

[0191] Table 1

[0192]

[0193] Scenario 1: Abnormal abrupt change in measurement:

[0194] In this scenario, the measurement data of cooperative navigation (i.e., the relative measurement values ​​between the lead aircraft and the wingman) will be subject to external interference and produce anomalies. Assume that there is a 10m ranging abrupt change during the ranging process in the following three time periods (630s, 650s), (930s, 950s), and (1230s, 1250s).

[0195] Figure 6 The root mean square error (RMSE) of the wingman's position obtained by UKF, UKF-ANFIS, and UKF-DEANFIS under the condition of abrupt measurement changes is presented. It can be seen that within the time period without abrupt measurement changes, all three algorithms can estimate the wingman's position information relatively accurately. However, in the time periods of (630s, 650s), (930s, 950s), and (1230s, 1250s), the performance of traditional UKF is severely affected by the impact of abrupt measurement changes, and its position accuracy decreases significantly. In contrast, UKF-ANFIS and UKF-DEANFIS can significantly mitigate the impact of abrupt measurement changes, obtaining smaller position errors. Figure 7 The cumulative root mean square errors (RMSEs) of the wingman positions obtained by UKF, UKF-ANFIS, and UKF-DEANFIS during periods of anomalous measurement changes were compared intuitively. It can be seen that UKF has the worst positioning accuracy, with RMSEs of 4.35m, 2.18m, and 0.29m, respectively. Compared to UKF, UKF-ANFIS and UKF-DEANFIS exhibit stronger robustness in suppressing anomalous measurement changes. Furthermore, the estimation accuracy of UKF-DEANFIS is very close to that of UKF-ANFIS. The RMSEs of UKF-ANFIS are reduced by 85.98%, 88.53%, and 51.72% compared to UKF, respectively; while the RMSEs of UKF-DEANFIS are reduced by 85.52%, 87.16%, and 44.83% compared to UKF, respectively, both demonstrating stronger robustness.

[0196] Scenario 2: Non-Gaussian measurement noise:

[0197] To address this situation, a Gaussian mixture distribution is used to describe non-Gaussian noise. This means that the original Gaussian distribution of the measurement noise is interfered with by another Gaussian distribution, resulting in non-Gaussian measurement noise.

[0198] ;

[0199] In the formula, The disturbance parameter representing noise pollution is set to 0.3 in this case. This represents the original Gaussian distribution, while This represents a perturbed Gaussian distribution with a standard deviation 10 times that of the original distribution. In this case, it is assumed that the measurement noise is contaminated by a perturbed Gaussian distribution over a time interval of (600s, 1200s).

[0200] The root mean square error of the position obtained by UKF, UKF-ANFIS and UKF-DEANFIS is as follows: Figure 8 As shown, within the range of (600s, 1200s), due to the influence of the perturbed Gaussian distribution, all three filters deviated from their original error magnitudes. Among them, the UKF filter exhibited a significantly increased navigation and positioning error due to its poor ability to suppress non-Gaussian noise interference. In contrast, UKF-ANFIS and UKF-DEANFIS utilized designed prediction and optimization adjustment strategies to improve estimation accuracy, effectively overcoming the impact of non-Gaussian measurement noise on navigation solutions. Furthermore, Figure 9 The cumulative root mean square error of position was intuitively compared using the three algorithms mentioned above. Compared with UKF, UKF-ANFIS and UKF-DEANFIS reduced latitude errors by 25.58% and 23.26%, respectively, longitude errors by 21.05%, and altitude errors by 28.57% and 21.43%, respectively. UKF-DEANFIS and UKF-ANFIS exhibited similar navigation estimation accuracy performance, indicating that UKF-DEANFIS retains the strong adaptability and robustness of UKF-ANFIS and can effectively suppress the negative impact of non-Gaussian measurement noise.

[0201] Scenario 3: Missing Measurements

[0202] During drone swarm flight, measurement data from wingmen may be missing due to obstacle obstruction or lead aircraft maneuvering. To simulate this situation, assume that no measurement data is available at time intervals of (600s, 650s) and (1200s, 1300s).

[0203] Figure 10 The root mean square error of the position calculated by UKF, UKF-ANFIS and UKF-DEANFIS in this case is given. Figure 11 A direct comparison of the cumulative root mean square error (RMSE) of the three filtering algorithms used for wingman position estimation is presented. It can be observed that during time intervals without available measurement data, the wingman can only rely on a low-precision MIMU for self-localization due to the lack of usable data, resulting in significant navigation and positioning errors. Figure 10The positioning results from UKF showed divergence. In contrast, UKF-ANFIS and UKF-DEANFIS utilize MIMU information and the lead aircraft's navigation data as input to the ANFIS prediction model to predict measurement data, effectively mitigating the navigation error divergence problem caused by the lack of measurement data. Compared to UKF, UKF-ANFIS and UKF-DEANFIS reduced latitude errors by 82.15% and 81.96%, respectively; longitude errors by 93.93% and 93.82%, respectively; and altitude errors by 81.01% and 78.48%, respectively. These results demonstrate that the proposed UKF-DEANFIS possesses strong predictive capabilities, effectively suppressing the adverse effects of missing measurement data and providing reliable measurement predictions for the MIMU to avoid navigation error divergence.

[0204] (2) Real-time performance evaluation:

[0205] Based on the three typical outlier measurement scenarios described above, Monte Carlo simulations were used to evaluate the proposed UKF-DEANFIS in terms of computational and real-time performance compared to UKF and UKF-ANFIS. The computing platform was equipped with Intel® Core processors. TM On a computer with an i9-12900H 2.5GHz processor and 16GB of RAM, a Matlab program was used to perform Monte Carlo simulations 20 times.

[0206] To eliminate the influence of different computer and processor performance, the relative computational efficiency of UKF-ANFIS and UKF-DEANFIS relative to UKF was calculated, and... Figure 12A direct comparison was made. It can be seen that the UKF, as a standard filter, requires the shortest computation time for navigation calculation. However, it lacks robustness in handling anomalous measurements. The computation time of UKF-ANFIS is almost twice that of UKF (163.34%, 181.49%, and 145.02% respectively in the three cases mentioned above), significantly exceeding the computation time of UKF and making real-time performance difficult. Meanwhile, the proposed UKF-DEANFIS reduces computation time by 54.59%, 54.96%, and 24.97% compared to UKF-ANFIS, respectively, very close to that of UKF. This is because UKF-ANFIS implements its designed ANFIS-based measurement prediction and measurement noise covariance optimization strategy at each time step, introducing a huge computational burden. However, UKF-DEANFIS utilizes a dual-event triggering mechanism to avoid unnecessary computational burden during periods without anomalous measurements; it only executes its designed ANFIS-based prediction and optimization strategy when event conditions are triggered. Therefore, the proposed UKF-DEANFIS achieves better real-time performance while ensuring the robustness of ANFIS.

[0207] The simulation results regarding robustness and real-time performance evaluation show that the proposed UKF-DEANFIS not only retains the superior robustness of ANFIS and enhances the adaptive ability of traditional UKF to handle the influence of abnormal measurements, but also achieves better real-time performance by constructing a dual-event triggering mechanism compared to UKF-ANFIS, thereby improving the application performance of UAV swarm cooperative navigation in complex special environments.

[0208] The steps of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be split into multiple steps. As long as they include the same logical relationship, they are all within the protection scope of this invention. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, without changing the core design of the algorithm and process, are also within the protection scope of this invention.

[0209] Another embodiment of the present invention relates to a UAV swarm cooperative navigation and positioning system. The implementation details of this UAV swarm cooperative navigation and positioning system are described below. The following implementation details are provided for ease of understanding and are not essential for implementing this solution. The UAV swarm cooperative navigation and positioning system of this embodiment includes:

[0210] The model building module is used to build a measurement model of the UAV swarm collaborative navigation system in order to obtain the collaborative measurement data of the UAV swarm obtained by all navigation sensors in the UAV swarm collaborative navigation system.

[0211] The noise adjustment module is used to obtain the theoretical and actual innovation covariance of the unscented Kalman filter UKF when performing cooperative navigation and positioning based on cooperative measurement data, and to adjust the theoretical measurement noise variance in the theoretical innovation covariance so that the theoretical innovation covariance after adjusting the theoretical measurement noise variance is equal to the actual innovation covariance.

[0212] The data acquisition module is used to extract relative measurement data from the collaborative measurement data of the UAV swarm collaborative navigation system relative to the navigation sensor; wherein, the relative measurement data consists of the calculated relative position / velocity between the leader and wingmen in the UAV swarm and the relative position / velocity measurement values ​​between the leader and wingmen measured by the relative navigation sensor;

[0213] The position / velocity prediction module is used to capture the dependency between the calculated and measured values ​​of relative position / velocity in the relative measurement data, and predict the measured values ​​of relative position / velocity through the dependency and the calculated values ​​of relative position / velocity, so as to reconstruct the relative measurement data by combining the predicted measured values ​​of relative position / velocity and the calculated values ​​of relative position / velocity.

[0214] The navigation and positioning module is used to perform collaborative navigation and positioning of UAV swarms based on the UKF after adjusting the theoretical measurement noise variance, or to enable the UKF to perform collaborative navigation and positioning of UAV swarms based on the reconstructed measurement data.

[0215] It is not difficult to see that this embodiment is a system embodiment corresponding to the above method embodiments, and this embodiment can be implemented in conjunction with the above method embodiments. The relevant technical details and technical effects mentioned in the above embodiments are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the above embodiments.

[0216] It is worth mentioning that all modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. Furthermore, to highlight the innovative aspects of this invention, this embodiment does not introduce units that are not closely related to solving the technical problem proposed by this invention; however, this does not mean that other units are absent from this embodiment.

[0217] Another embodiment of the present invention relates to a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the UAV swarm cooperative navigation and positioning method of the above embodiments.

[0218] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and memories. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and will not be described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.

[0219] The processor manages the bus and general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory is used to store data used by the processor during operation.

[0220] Another embodiment of the present invention relates to a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the method embodiments described above.

[0221] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0222] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing the present invention, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of the present invention.

Claims

1. A method for cooperative navigation and positioning of unmanned aerial vehicle (UAV) swarms, characterized in that, The method includes: Establish a measurement model for a UAV swarm collaborative navigation system to obtain collaborative measurement data of the UAV swarm obtained through all navigation sensors in the UAV swarm collaborative navigation system; The theoretical and actual innovation covariance of the unscented Kalman filter (UKF) during cooperative navigation and positioning based on cooperative measurement data are obtained. The theoretical measurement noise variance in the theoretical innovation covariance is adjusted so that the theoretical innovation covariance after adjusting the theoretical measurement noise variance is equal to the actual innovation covariance. The relative measurement data is extracted from the cooperative measurement data of the UAV swarm cooperative navigation system relative to the navigation sensor; wherein, the relative measurement data consists of the calculated relative position / velocity between the leader and wingmen in the UAV swarm and the relative position / velocity measurement values ​​between the leader and wingmen measured by the relative navigation sensor; Capture the dependency between the calculated and measured relative position / velocity values ​​in the relative measurement data, and predict the measured relative position / velocity values ​​using the dependency and the calculated values, so as to reconstruct the relative measurement data by combining the predicted and calculated values. The UKF is used to perform collaborative navigation and positioning of UAV swarms based on the adjusted theoretical measurement noise variance, or the UKF is used to perform collaborative navigation and positioning of UAV swarms based on the reconstructed measurement data.

2. The UAV swarm cooperative navigation and positioning method according to claim 1, characterized in that, The adjustment of the theoretical measurement noise variance in the theoretical innovation covariance to make the adjusted theoretical innovation covariance equal to the actual innovation covariance includes: The theoretical information covariance and the actual information covariance are input into the first adaptive neurofuzzy inference network ANFIS; The matching degree between the theoretical innovation covariance and the actual innovation covariance is obtained through the ANFIS network, and an adjustment coefficient is output based on the matching degree to adjust the theoretical measurement noise variance so that the theoretical innovation covariance after adjusting the theoretical noise variance is equal to the actual innovation covariance.

3. The UAV swarm cooperative navigation and positioning method according to claim 1, characterized in that, The process of capturing the dependency between the calculated relative position / velocity value and the measured relative position / velocity value in the relative measurement data, and predicting the measured relative position / velocity value using the dependency and the calculated relative position / velocity value, includes: Using the calculated and measured relative positions / velocities of the lead and wingmen of several drone swarms as sample data, a second ANFIS network is trained to obtain an ANFIS model, which captures the dependency between the calculated and measured relative positions / velocities. The calculated relative position / velocity values ​​are input into the ANFIS model to obtain the predicted relative position / velocity measurements.

4. The UAV swarm cooperative navigation and positioning method according to claim 1, characterized in that, Before obtaining the theoretical and actual innovation covariance of the unscented Kalman filter (UKF) during cooperative navigation and positioning based on cooperative measurement data, the method further includes: Construct an anomaly detection function for collaborative measurement data; The anomaly detection function is: ； In the formula, It is a unit vector, and the first... Each element is 1. It is the dimension of the measurement vector. For the innovation vector, This is the corresponding covariance matrix; Based on the anomaly detection function, set the first event trigger condition, and determine whether the collaborative measurement data is abnormal based on the first event trigger condition; The trigger condition for the first event is: ； In the formula, It is the false alarm rate. It is the detection threshold.

5. The UAV swarm cooperative navigation and positioning method according to claim 4, characterized in that, Before obtaining the theoretical and actual innovation covariance of the unscented Kalman filter (UKF) during cooperative navigation and positioning based on cooperative measurement data, the method further includes: If the collaborative measurement data is determined to be abnormal, then a relative position change index of the collaborative measurement data is constructed. The relative position change index is: ； In the formula, Indicates the measurement dimension involved in the anomaly. The total dimension of the measurement. It is during the quarantine period After the dimensional measurement, the 3rd dimension of the state estimation error covariance matrix diagonal elements, During UKF filtering, the state estimation error covariance matrix is ​​the first... One diagonal element; Based on the relative position change index, set the second event trigger condition, and based on the second event trigger condition, determine whether to adjust the theoretical measurement noise variance or whether to predict the relative position / velocity measurement value. The trigger condition for the second event is: ； In the formula, It is the threshold for triggering the second event.

6. A collaborative navigation and positioning system for unmanned aerial vehicle (UAV) swarms, characterized in that, The system includes: The model building module is used to build a measurement model of the UAV swarm collaborative navigation system in order to obtain the collaborative measurement data of the UAV swarm obtained by all navigation sensors in the UAV swarm collaborative navigation system. The noise adjustment module is used to obtain the theoretical and actual innovation covariance of the unscented Kalman filter UKF when performing cooperative navigation and positioning based on cooperative measurement data, and to adjust the theoretical measurement noise variance in the theoretical innovation covariance so that the theoretical innovation covariance after adjusting the theoretical measurement noise variance is equal to the actual innovation covariance. The data acquisition module is used to extract relative measurement data from the collaborative measurement data of the UAV swarm collaborative navigation system relative to the navigation sensor; wherein, the relative measurement data consists of the calculated relative position / velocity between the leader and wingmen in the UAV swarm and the relative position / velocity measurement values ​​between the leader and wingmen measured by the relative navigation sensor; The position / velocity prediction module is used to capture the dependency between the calculated and measured values ​​of relative position / velocity in the relative measurement data, and predict the measured values ​​of relative position / velocity through the dependency and the calculated values ​​of relative position / velocity, so as to reconstruct the relative measurement data by combining the predicted measured values ​​of relative position / velocity and the calculated values ​​of relative position / velocity. The navigation and positioning module is used to perform collaborative navigation and positioning of UAV swarms based on the UKF after adjusting the theoretical measurement noise variance, or to enable the UKF to perform collaborative navigation and positioning of UAV swarms based on the reconstructed measurement data.

7. A computer device, characterized in that, include: At least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the UAV swarm cooperative navigation and positioning method as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the UAV swarm cooperative navigation and positioning method as described in any one of claims 1 to 5.

Images